Since the first immune checkpoint inhibitor was approved to treat advanced melanoma in 2011, impressive improvements in clinical outcomes have continued to be demonstrated across several cancer types. However, not all patients benefit from these agents and many studies have focused on identifying predictive and prognostic biomarkers in an attempt to better inform and guide treatment decisions.
Several studies investigating established and novel biomarkers of response to immunotherapy will be presented at this year’s Congress.
Programmed death ligand-1 (PD-L1) expression has been one of the most hotly debated biomarkers in immuno-oncology since the introduction of PD-1/PD-L1 immune checkpoint inhibitors. There are currently multiple assays under investigation or approved as companion or complementary diagnostic tests for PD-L1 expression. It is a concern that the US FDA approval of an assay on the basis of its performance appears to have become more important than the accurate and reproducible measurement of the target. As a result, at least four separate antibodies have been included in assays that are part of separate FDA submissions, creating a challenge for pathologists who may need to perform four different assays rather than simply assess PD-L1 expression. Indeed, a recent comparative study found that differences reported in PD-L1 expression in lung cancer tissue arose from tumour heterogeneity or the assay or platform used, rather than the choice of antibody.1 Imagine a situation where a pathologist was required to use separate assays to assess the dozen or so drugs that target the oestrogen receptor in breast cancer.
Two Late-Breaking Abstract presentations will describe efficacy data in advanced non-small-cell lung cancer (NSCLC) by PD-L1 expression status. In the first, overall survival data will be presented from the first phase III study of atezolizumab versus docetaxel (Abstract LBA44_PR), while in the second, preliminary efficacy data will be presented from the first study to combine anti-VEGF (ramucirumab) and anti-PD-1 (pembrolizumab) antibody treatments (Abstract LBA38). The value of PD-L1 expression as a biomarker of response in melanoma is also considered in a pooled analysis of phase II (CheckMate 069) and phase III (CheckMate 066 and 067) trials comparing nivolumab plus ipilimumab versus nivolumab alone. Data in advanced melanoma appear to be far from clear-cut and PD-L1 expression does not seem to predict response to immune-targeting drugs (Abstract 1112PD).
Multiple diagnostic assays are available for determining PD-L1 expression status and have been used in clinical trials of different immunotherapies. Data from a study comparing three PD-L1 diagnostic assays from biopsies of squamous cell carcinoma of the head and neck (SCCHN) show a strong correlation between the assays, suggesting that it may be feasible to compare data derived from different PD-L1 diagnostic tests (Abstract 955PD).
To date, a number of regulatory approvals for PD-(L)1-targeting agents are linked to companion diagnostic assays and there are potential risks associated with cross-matching agents to assays in the absence of established clinical and analytical concordance, according to Dr Jorge Martinalbo from the European Medicines Agency, London, UK. Further confusion comes from different scoring criteria and thresholds for defining PD-L1 positivity, which vary by agent and tumour type. Acknowledging that harmonisation of assays is probably unrealistic, a blueprint proposal initiative was started in 2015 with the remit to ‘agree and deliver, via cross-industry collaboration, a package of information/data upon which analytic comparison of the various diagnostic assays may be conducted, potentially paving the way for post-market standardisation and/or practice guideline development, as appropriate’.2
In patients with advanced cancer, the relationship between tumour mutational burden and microsatellite instability both appear to be of value in identifying patients most likely to derive benefit from immunotherapy (Abstract 52O). However so far, there is no single reliable, validated biomarker for selecting patients who are likely to benefit from immunotherapies. At the moment PD-L1 expression, CD8+ T-cell infiltrates and ‘foreignness’ of the tumour, despite all being correlated with response or survival to immunotherapy with checkpoint inhibitors, are not sufficiently robust to discriminate with high specificity and sensitivity between those patients who would and would not benefit. The reason for this could be that different treatment-evasive mechanisms may play a role across tumour types. This is quite different for targeted agents that require a specific gene mutation or translocation in order to be active. In particular, because the overall response rate to immunotherapy for many tumour types is modest, improved selection criteria are becoming more urgent as we expose our patients to sometimes highly toxic drugs. Establishing predictive biomarkers is also becoming increasingly important from a health economic perspective. The cost of immunotherapies is such that it impacts ever more on the total healthcare budget, which in turn affects the availability of these drugs in different European countries. We at the Netherlands Cancer Institute have developed the ‘cancer immunogram’, an initial framework of seven parameter classes describing cancer–immune interactions for individual patients.3 This may become a tool to help oncologists assess the likelihood of benefit from immunotherapy in the future.
- Gaule P, et al. JAMA Oncol 2016. Aug 18. Epub ahead of print
- www.fda.gov/downloads/MedicalDevices/NewsEvents/WorkshopsConferences/UCM439440.pdf
- Blank CU, et al. Science 2016;352:658–60
This article appeared in the Saturday edition of the Daily Reporter