Predicting immunotherapy response to maximise benefit for all patients

The topic of cancer immunotherapy is as hot as ever. It was in the spotlight of many of the sessions at the annual meeting of the American Association of Cancer Research (AACR) in Chicago this year. This is one of the biggest cancer conferences in the world, with a record 22,000 people attending this year. Together with 2 fellow MELGEN PhD students, Joanna Pozniak and Sathya Muralidhar, I had the opportunity to attend and present my work. Thanks to MELGEN and a generous travel grant from Newnham College, University of Cambridge.

The idea at the forefront of cancer immunotherapies is targeting the interaction between PD-1 and PD-L1. This shows impressive results in the clinic for patients with advanced melanoma as well as other cancer types. By blocking this interaction, it’s like releasing the brakes on the immune system – which unleashes this force of tumour destroyers. Several such immunotherapy drugs, including PD-1 inhibitors: Pembrolizumab (Keytruda®), Nivolumab (Opdivo®) and PD-L1 inhibitors: Atezolizumab (Tecentriq®), Durvalumab (Imfinzi®), Avelumab (Bavencio®), are currently approved for various cancers

Over the 5 conference days, one of the major challenges discussed were the difficulties in predicting response to immunotherapy. In an ideal world, every drug that is ever created would work equally well for every patient. Sadly, this is not true (we are all different), and the best we can do is try to find ways of figuring out who is most likely to respond to a treatment. Some of these ways are as follows:

PD-L1 expression:
Several studies propose PD-L1 expression as a biomarker of response to immunotherapy. The hypothesis is, that tumours with high expression of PD-L1 are more likely to respond to therapies targeting the PD-1/PD-L1 interaction. In reality, it is much more complicated. There is an association between higher PD-L1 expression and response to immunotherapy. However, patients with low expression of PD-L1 can also respond. Hence a decision on whether a patient should receive immunotherapy based solely on PD-L1 expression is very difficult to make.   

Tumour mutational burden (TMB):
Recent work has suggested using tumour mutational burden as a predictive biomarker. Patient with a higher number of mutations in their tumours – are more likely to respond to immunotherapy. I briefly described one possible reason behind this hypothesis in a previous blog post (basically, more mutations -> more likely to awaken the immune response). But again, TMB is not a perfect biomarker as patients with low TMB still can have a good response to immunotherapy. One explanation for this is that such patients can have few but very immunogenic mutations. This gives rise to immune cell recognition and attack when receiving immunotherapy drugs.

IFN-γ pathway genes:
In short, IFN-γ is a pro-inflammatory cytokine, secreted by immune cells. High levels of IFN-γ in the tumour microenvironment suggests presence of immune cells, which can become activated through immunotherapy. A number of studies have linked high expression of IFN-γ-associated genes to favourable immunotherapy response. Loss of function mutations in genes linked to the IFN-γ pathway have also been reported to cause lack of response to immunotherapies.



There are many more potential markers of immunotherapy response being researched. I have highlighted a few in this post. If you’re interested in finding out more – there is an excellent short summary published in Science. In conclusion, there are several promising biomarkers already. But, using them in isolation is an imperfect approach. Clinical research would greatly benefit from using combinations. Additionally, we need more novel biomarker research as well – which is the topic of some PhD projects as part of the MELGEN consortium.

Written by Sofia Chen, ESR07