By Mike Birtwistle and Michael Speer
Traditionally, most cancer treatments have been blunt instruments. We have sought to cut out tumours, zap them with radiation or poison them with chemotherapy. In doing so, we have known little about the individual characteristics of the patient or their tumour, including what made it grow (or not). Treatment has been delivered at scale, on a patient population, rather than a personal, level.
The excellent Emperor of All Maladies by Siddhartha Mukherjee details how our understanding of cancer has evolved – radically – over the years. In the past, treatment was too often based on intuition rather than understanding. Sometimes this proved to be remarkably successful, but all too often it did not. The end result was that too many patients were exposed to unnecessary or ineffective treatment, often at the cost of great side effects.
This has all begun to change. The days of hitting patients with ever more invasive and toxic combinations of treatment are over. Thanks to scientific advances, especially in genetic profiling, we now know that most forms of ‘cancer’ are in fact not one disease, but a combination of often very different cancers, with very different behaviours, which just happen to be located in a particular part of the body. We know this to be true for breast, bowel, lung, skin, prostate and blood cancers. It is almost certainly true for every other tumour type as well. We also know that each cancer can be driven by a few significant genetic changes.
One of the most notable examples of the potential of targeted medicines has been provided by the Nobel Prize winner Professor Ralph Steinman, who lived for over four and a half years following a diagnosis of pancreatic cancer – a hideous disease that sees most patients survive less than a year. Professor Steinman used cells that were part of his immune system, dendritic cells, which he had originally discovered through his research back in the 1970s. He was able to train these cells to identify and attach themselves to his cancer cells, flagging them up to be attacked by other cells that make up the body’s immune system, B and T cells. By working with a global network of colleagues he brought together different strands of science to effectively use his own immune system to create a vaccine to target his cancer cells.
This is personalised treatment in the extreme and, for most patients, this is some way off. However, we are now able to stratify cancers.
Treatments for breast cancer have become the poster girls for this new understanding. We know that some forms of breast cancer are receptive to hormones, whereas others are not. We also know that some breast cancers over-express different biological markers, such as HER2, and that drugs can be developed to target this, transforming the prospects for some types of breast cancer which were historically seen as being stubbornly immune to treatment. In the case of HER2-positive breast cancer, we already have Herceptin and Tyverb as treatments targeted at this sub-group of patients, with the promise of many more drugs which will work in combination with these to come.
Yet we also know that other tumours – such as stomach cancer – can also over-express HER2. Perhaps, rather than having cancers of different parts of the body which might have different biomarkers, we actually have cancers linked to different biomarkers, which might then occur in different parts of the body?
Although it makes sense (for obvious reasons) for cancer surgeons to specialise in particular parts of the body, perhaps future generations of oncologists will specialise in treating, for example, HER2 positive cancer. Maybe even patients will begin to identify as much with their genetic mutation as they do with the location of their cancer. When, how and in what order to use targeted treatments will require particular expertise. The development of targeted medicines poses some interesting questions for the way we think about cancer and organise services to treat it.
Although the HER2, KRAS, EGFR and BRAF markers are now used as important indicators for treatment, there will be many more such markers to come. We probably already know about many of these other markers, but just do not yet have the treatments to make them useful. This could all change in a short space of time and there is an opportunity for the UK to lead this change. A national health service, high quality cancer registration and world class research are all advantages which the UK should be well-positioned to exploit – a point made in the recent life sciences strategy.
Cancer Research UK has recently announced its plans to recruit up to 9,000 patients in its stratified medicines initiative, which aims to test cancer tumours for a wide range of biomarkers, hopefully identifying patterns in genetic mutations which will offer clues to future treatments, fast tracking the development of new drugs. It also offers hope for some drugs which previously failed to show efficacy in trials. Perhaps, upon re-examination, these drugs will prove to be effective for some patients with particular biomarkers? This has already happened with a lung cancer drug whose efficacy in EGFR positive patients was hidden in a wider trial. Intriguingly, some clinical trials are already beginning where the primary entry criterion is a genetic characteristic rather than the location of a tumour.
This vision of the future is exciting in many respects. It could focus particular treatments on the patients who are most likely to benefit, enabling clinicians to recommend each patient the treatment and dose which is most likely to work for them, rather than simply treating according to a prescribed pathway. Stratified medicine also offers the hope of accelerating the process of developing and testing new drugs – something which both patients and pharmaceutical companies would welcome.
However, if this vision is to be realised, then there are significant challenges to be overcome. Testing for cancer biomarkers requires specialist molecular diagnostics. This in turn requires training, strict calibration of test results and investment in equipment. Simply allowing every pathology service to dabble in this will not work. Specific providers will need to be designated, ensuring they process sufficient volumes to build up specialist expertise. Other pathology services will need to forge close links with these specialist hubs, preventing molecular diagnostics becoming the domain of a few ivory towers. The NHS will need to learn from the experience of Cancer Research UK and those pharmaceutical companies who have already supported the development of some diagnostic capacity. Equally, commissioners will need to make available the resources to fund tests, and the links between routine practice and research will need to be strong.
Personalised treatments offer obvious benefits to patients, as the treatments are more effective for their cancer, dosing levels can be calculated more accurately, and patients whose genetic make-up means they will not benefit can be spared the possible side-effects. There are also clear benefits for the NHS as the medicines that are used will be the ones with the greatest value and patient benefit, and treatments that will not be effective do not have to be tried, saving money and capacity. Personalised testing also has the potential to benefit prevention and early diagnosis by identifying patients with a genetic risk of getting a particular cancer, offering the potential for preventative action, for example by modifying their lifestyle and through dedicated monitoring by the NHS.
Targeted treatments will also change both the business model for pharma and the way in which health services will need to approach these drugs. The days of the cancer blockbuster are probably behind us. Instead, treatments will be focused on ever smaller groups of patients. Breast cancer will not be one disease affecting over 48,000 people a year in the UK, but tens of diseases affecting groups as small as 1,000 patients, which will have implications for the way in which clinical trials are organised, as well as the resultant evidence of clinical benefit which is deemed to be acceptable. Although this could reduce overall health service costs by ensuring that fewer patients receive the drug, it will necessitate higher unit costs, as manufacturers will need to make a return on their investment. Health technology assessment processes will have to adapt to reflect this – another challenge for value-based pricing.
For the UK to realise its potential as a leader in this field, we also need to ensure that patients are able to receive the latest treatments in routine NHS care. Otherwise we risk ruling ourselves out as a destination for important clinical trials, as most patients will not have been treated with the existing standard of care, which is usually a pre-requisite for entry into trials. In this respect the Cancer Drugs Fund could have an important (and partly unintended) benefit (so far, nearly 10,000 patients have accessed treatments which would otherwise have had little NHS usage – an important resource for future trials).
Personalised medicine in the form practised by Professor Steinman is not even on the horizon for routine care. But treatment targeted at the personal characteristics of a patient is already here and will become the norm, rather than the exception, over the next decade. This creates great opportunities, but also challenges. The future of cancer is increasingly personal.