Professor Gary Middleton, chair of Medical Oncology at the University of Birmingham,
is the national lead for a clinical trial of personalized treatment for lung cancer termed The National Lung MATRIX. The trial forms the second phase of Cancer Research UK’s Stratified Medicine Programme. In this month’s Clinical Feature, Gary explains the thinking behind the MATRIX trial, and outlines how it provides a basis for testing promising new immunotherapies on patient subgroups most likely to benefit from them.
Mastering the MATRIX
By Gary Middleton
Despite the fact that lung cancer is extremely heterogeneous, with many different subtypes associated with distinct genetic, biological, and clinical properties, until 10 years ago, every patient by and large received the same treatment. Most depressingly, treatment for non-small cell lung cancer, which makes up ~80-85% of all lung cancers was restricted to use of chemotherapy regimens which had limited efficacy, but were highly toxic, with side effects such as nausea and vomiting, alongside damage to the bone marrow.
The situation started to change around 2004, when inhibitors of the Epidermal Growth Factor Receptor tyrosine kinase were found to provide clinical benefit to lung cancer patients, but only in patient subgroups bearing certain genetic mutations in the receptor tyrosine kinase domain. These mutations were the key driver of oncogenesis in these patients and could be switched off with the inhibitor. This was the first indication that lung cancer patients could be usefully “stratified” into separate groups, based on the different genetic mutations in their tumours, some of which might respond more effectively to particular therapeutic drugs. Since then, a number of novel genetic mutations have been defined for subgroups of lung cancer patients, each now associated with a different therapy specifically targeted at patients bearing the relevant mutation.
This new approach represents a major shift in how to treat cancer generally – a shift from generic treatments applied across a disease group to personalized treatments focussed on the particular characteristics of a patient’s tumour – ie personalized medicine. The obvious benefit to patients is in an increased chance of therapeutic benefit, and decreased side effects resulting from a more focussed application of new agents.
The MATRIX trial that I lead is the most ambitious application of the personalized medicine philosophy to any cancer treatment – certainly in the UK, and probably, internationally. The trial leads on from phase 2 of the UK’s Stratified Medicine Programme (SMP2), and will stratify around 2000 patients per year diagnosed with stage III or stage IV lung cancer into an unprecedented number of subgroups, based on which genetic mutations are present in the tumour, and then treat them with the appropriate targeted therapy for their genetic abnormality. This will create a unique “matrix” defining numerous patient subgroups, each of which is potentially the recipient of new “targeted” therapies based on the particular mutations present. The unprecedented size of the trial is important because it means each molecular subgroup should have enough patients in it to make testing individual targeted therapies feasible – and statistically meaningful.
The Matrix Trial will therefore operate as an “umbrella” under which will sit multiple individual (Phase II) trial arms, each testing in parallel an experimental drug in a population stratified by multiple pre-specified target genetic “biomarkers”. In many cases, the trials won’t rely on previous targeted therapies – and the combinations of specific drugs used, and subpopulation treated, will be completely novel. It is my job to coordinate this overall strategy – and it is one which will involve multiple centres, and for which collaboration with industry, the source for many of the therapeutics to be used in Matrix, will be critical.
Although initially many of the drugs being tested in the MATRIX will be targeted therapies aimed at specific molecular targets there will be a strong emphasis on trialling novel immunotherapies. 2013 was hailed by Science magazine as a breakthrough year for cancer immunotherapy, and it is not difficult to see why. One of the most dramatic changes has been the advent of checkpoint blockade therapy – this is the idea, pioneered by Prof Jim Allison’s group in the US, of using antibodies to block inhibitory signals transmitted to the patient’s T cells via receptors on their surface, such as CTLA-4 or PD-1. This strategy effectively “removes the brakes” from the immune response, potentially unleashing strong anti-tumour immunity.
The results, from trials in melanoma (where it is now approved as a treatment) and other tumours including lung, have been remarkable: in a sizeable minority of patients, the therapy appears to induce durable remissions. And as shown in the pictures below, I have started to see similarly remarkable tumour regressions in some of my own patients. However, in some cases removing the brakes on T cell reactivity is accompanied by the immunological equivalent of a car crash – an unleashing of strong T cell responses against the patient’s normal tissues, resulting in autoimmune-like symptoms that have in extreme cases proven fatal. And in many other patients, the checkpoint blockade strategy is simply ineffective.
Determining which group of patients checkpoint blockade immunotherapy is likely to be effective in is critical – and the suspicion is this will relate to particular features of the tumour microenvironment in the patient. Are there T cells present that can recognize the tumour ? Do they express relevant inhibitory receptors ? Does the tumour express relevant targets for T cell recognition ? These questions are currently hot topics in the field, and the research my laboratory is carrying out is aiming to address these issues.
By testing efficacy of these therapies in Matrix, it should allow us to start to see where molecular genetics meets immunology, by correlating three key factors: the molecular phenotype of particular patient subgroups, the response to treatment, and also the nature of the tumour microenvironment. Potentially, simple genetic biomarkers might emerge from Matrix that accurately predict responsiveness to such immunotherapies – which could ultimately affect future choice of standard treatments for lung cancer, and by focussing treatments on certain subgroups most likely to benefit from them, would benefit both patients themselves, and those paying for these expensive immunotherapies.
If and when this happens, it will be a sign we are starting to master the Matrix.
To read or view more about the MATRIX trial, check out the following links: