The arrival of immunomodulatory agents has led to a paradigm shift in oncological practice. Immune checkpoint blockers such as nivolumab and ipilimumab have become standard of care in many malignancies, offering good chances of durable responses. The field of immuno-oncology is expanding rapidly and many dozens of immuno-oncological agents targeting distinct molecular pathways are currently in clinical trials.
Significant limitations remain, however: severe immune-related toxicity affects more than 50% of patients receiving combination ipilimumab/nivolumab regimens; a minority of patients may experience ‘hyperprogression’ in response to immunotherapy; the costs of these agents are not insignificant despite response rates in most tumour types being in the region of 15–30% for single agents.
Using combination regimens may improve efficacy, but in practical and financial terms it is not feasible to test each possible combination for every indication in unselected groups of patients. Preclinical studies to narrow down the field are challenging; the human immune system has marked differences even from that of closely related primate species, and animal models have repeatedly proven inadequate to predict clinical efficacy and toxicity.
Being able to predict which patients are more likely to respond to specific immunomodulatory approaches and how likely they are to experience specific side effects would alleviate these issues. This is the holy grail of current research in the field of immuno-oncology.
Challenges in developing biomarkers for tumour immunotherapy
Attempts at empirically discovering measurable indicators using peripheral blood have thus far failed to deliver usable biomarkers. One likely reason is that the crucial immune events take place in or adjacent to the cancer tissue. The picture derived from peripheral blood may bear very little resemblance to what is happening in the cancer tissue and may completely miss crucial cell populations such as tissue-resident memory T cells. Additionally, identifying peripheral blood-based proxy markers for events in the tumour microenvironment is extremely challenging without a good understanding of the underlying immune mechanisms.
To rectify this problem, access to suitable biopsy samples at critical time points is essential. This is not trivial to arrange and until recently was not routinely part of clinical trial design; moreover, in many tumour types access to tumour tissue is difficult at best and poses additional risks of biopsy-related discomfort and morbidity.
Recent developments in genomic, transcriptomic and proteomic techniques are rapidly expanding our knowledge of individual cell types, molecules and pathways involved in the human immune system. At the same time they expose critical gaps in our understanding of the complex networks involved. Despite advances in bioinformatics, building a complete model is beyond our current capabilities and limits our ability to rationally design predictive biomarkers for use of immuno-oncological agents.
Finally, enthusiasm among pharmaceutical companies has led to the parallel development of multiple immuno-oncological agents targeting the same pathways with matched diagnostic tests. Unfortunately, standardization is lacking in ...