Is Precision Immunology Possible?

These approvals have resulted in challenges for the clinical oncologist in deciding which agents to use in which patients. While ultimately clinical studies must be done to determine the benefits and potential pitfalls of various drug sequences, a high priority has been placed on the use of predictive biomarkers to better select appropriate therapy for individual patients. If such biomarkers were available, it may be theoretically possible to determine the optimal immunotherapy drugs, combinations, and sequences for each patient. This would be similar to using genomic mutation profiles in cancer cells to help select targeted therapy, as established in so-called “precision medicine” clinical pathways.

I recently coined the term “precision immunology” to describe the use of host, immune system and tumor factors as biomarkers to select immunotherapy approaches. But, are we ready to integrate precision immunology into clinical practice?

There has been an extensive effort to evaluate a number of putative soluble factors, immune cell characteristics, and genomic immune signatures in clinical tumor immunotherapy trials. To date, none have achieved validation through prospective clinical studies. Nonetheless, several biomarkers appear to be gaining considerable attention.

The elegant work of Jérôme Galon, PhD, and colleagues have suggested that the number and location of immune cells, most notably CD3+ and CD8+ T cells, have prognostic value in patients with colorectal cancer. This finding is currently being prospectively validated in a global multi-institutional trial. The Cytokine Working Group recently completed a prospective study of several peripheral blood and tumor based biomarkers in melanoma patients undergoing treatment with high-dose IL-2.

Several clinical trials have reported an association between clinical benefit and tumor cell expression of the programmed cell death ligand 1 (PD-L1), the ligand for PD-1. The potential use of PD-L1 expression as a biomarker has been further fueled by the recent FDA approval of pembrolizumab, an anti-PD-1 agent, in lung cancer only for PD-L1-positive patients.

It may be important to emphasize that the FDA approval for pembrolizumab was based on the clinical trial, which used high PD-L1 expression as an eligibility criterion for study participation, and not on validation of PD-L1 as a predictive biomarker of response. This has complicated the field, because nivolumab, another anti-PD-1 agent, was approved for the treatment of non—small cell lung cancer without the need for pre-treatment PD-L1 testing, since this was not a criterion for inclusion in the nivolumab clinical trials.

Recently, there has been considerable excitement over the observation that patients whose tumors harbor a high mutation burden or are characterized by significant genomic instability, may respond better to T cell checkpoint inhibitors. This finding is interesting, but requires further validation to determine if the genomic changes are producing neoantigens capable of eliciting tumor rejection. There are indeed numerous other peripheral blood and tumor-derived factors and panels being intensely investigated.

While the identification of several potential biomarkers has been reported, none have yet been fully validated as predictive of clinical responses to tumor immunotherapy. Further, it is unclear whether the markers discussed will be specific to individual drugs, pharmacologic classes, or tumor types. So, how should we proceed?

The next few years will be a critical time to further evaluate the biomarkers currently showing promise while continuing to utilize advances in high-throughput technology and computational biology to help optimize the most promising biomarkers. Physicians should encourage their patients to participate in studies in which peripheral blood and tumor samples are collected to help confirm the clinical utility of potential biomarkers. These studies are increasingly important as more therapeutic options are being developed for patients.