Skip to main content

Together we are beating cancer

Donate now
  • Health & Medicine
  • Policy & Insight

Expert opinion: why are some cancers so difficult to treat?

by Oliver Childs | Analysis

26 January 2012

1 comment 1 comment

Can we chase tumours down evolutionary dead ends?

Professor Gerard Evan is Head of the Department of Biochemistry at the University of Cambridge. Earlier in the week, we wrote about his work on a crucial cancer-associated gene called Myc.

In this thought-provoking article, Professor Evan explains how cancer is an evolving and adapting enemy, and talks about strategies to combat the disease.

Biology has undergone an unprecedented technical revolution in the past two decades. Despite its complexity, biological systems can now be mapped and catalogued in minute detail – we can monitor the activity of every one of our approximately 25,000 genes; identify almost every protein present in a cell; and even sequence the entire genomes of animals, plants, bacteria or cancer cells.

For cancer researchers, these technologies can appear a Godsend: cancers are extremely diverse diseases that arise through the build-up of mutations (errors) in the genes that regulate and restrain the growth, division, and movement of the cells that make up our bodies. The process is ‘Darwinian’ – in other words, the mutations occur at random over our lifetimes, and the faulty cells then either die out or survive and multiply as a result of the complex, changing and still largely mysterious selective pressures in the body.

The technical revolution of recent years has undoubtedly advanced our understanding of cancer, and is helping sustain impressive improvements in cancer survival rates, but why don’t we hear the word ‘cure’ very often when it comes to this disease? And what needs to be done so that we do?

Mapping complexity

Not surprisingly, given the haphazard and random way in which they evolve, every patient’s cancer is unique. Indeed, mapping and understanding the complexity within just one person’s cancer could occupy an entire research institute. In some ways, cancer research has reached an existential impasse – we can map and catalogue and annotate forever, but what is it about cancer that we really want to know?

For most of us, I suspect the answer is simple and pragmatic: we want to know how to cure patients.

Simple questions have the habit of exposing others that are more fundamental: in this case, “Why are cancers so difficult to cure?” Here, there are two general schools of thought.

Personalised medicine

Many would point to the disconcerting genetic diversity of cancers – an inevitable consequence of the haphazard way that they evolve. In the past, cancer therapies were applied fairly indiscriminately, but now many believe that effective therapies need to be specific and tailored to the particular genetic faults in each individual’s cancer – in other words cancer therapy needs to be ‘personalised’. With the advent of new targeted drugs (such as Herceptin for some breast cancers, Glivec for certain leukaemias, and brand new drugs like vemurafenib for some skin cancers), many hope that this may, at least one day soon, be feasible.

But simply personalising treatment so that it targets the genetic faults present in a tumour at the point of diagnosis, disregards the most fundamental reason for why cancers are difficult to eradicate forever: cancer cells adapt and evolve in response to treatment.

Because of this, even drugs that are initially very effective often have a progressively dwindling effect over time, as the biological systems that are blocked by the treatment spontaneously compensate by re-routing the cancer cells’ internal wiring, thereby restoring the cancer’s ability to grow and spread. To use an analogy, traffic hot spots in towns can cause major traffic jams, but cunning drivers will quickly find short cuts to get round the congestion.

Then, in those rare situations where cancer cells cannot take such ‘short cuts’, evolution takes over: in response to drug treatment, spontaneously arising mutant cancer cells that are resistant to the targeted drug rapidly outgrow their incapacitated siblings and the cancer comes back.

Although some patients can be successfully cured and their cancers don’t return – such as is often the case for testicular cancer and some childhood cancers – for other cancers the situation is different, as so many patients and their families are all too aware. It doesn’t matter how effective or specific a therapy is – if the system the treatment is targeting can be bypassed by compensation or evolution, that therapy will become less effective over time and eventually fail, and the cancer will return.

Evolutionary dead end

Against two such formidable adversaries as compensation and evolution what are we to do? A solution is to identify targets that are essential for the survival of cancer cells but whose inhibition cannot be bypassed by compensation or evolution. Rather than causing localized traffic jams within the city that can be circumvented by short cuts, we identify the bridge that is the only way out of town and then block that.

But do such fundamental targets exist? Can we make drugs that inhibit them? How bad might the side effects of such therapies be, given that such essential and non-redundant engines of biology are likely to serve important functions in ‘normal’ bodily processes? We don’t yet know.

A second idea is to chase each cancer down an evolutionary valley and into a dead end from which it cannot escape. We accept that localized traffic jams can be bypassed but, each time that happens, we identify the back route and then target that – and we keep doing it until there are no short cuts left. In practice, this would mean treating a patient with one targeted drug and, if their cancer returns with newly-developed resistance to this treatment, we then identify how that resistance evolved and hit the tumour with another drug directed at that resistance mechanism. The process is repeated until the cancer runs out of evolutionary headroom.

Both of these strategies will need a significant re-tooling and refocusing of our cancer research enterprise. We need to pay more attention to the inherent robustness and evolutionary ability of the targets against which our drugs are directed, and see if we can identify the essential engines of tumour maintenance.

One potential benefit of finding such fundamental engines is that they might be a common requirement of many, perhaps all, cancers: in which case a therapy targeted against them might usher in an exciting and far more affordable era of ‘impersonalised’ cancer treatment.

On the other hand, if we are going to chase tumours into extinction, we need to track the evolutionary trajectories of cancers in patients through sequential episodes of treatment and relapse. To do this, we must establish what information we need and how to get it without putting patients through repeated and onerous biopsies and procedures.

Common goal

All cancer researchers share the common goal of wanting cancer therapies that are effective and durable. But to do this we must never lose sight of the fact that cancers are just another example of evolution at work.

Cancers are blind. They are neither clever nor cunning – but humans are.

We can – and will – beat them.


Professor Gerard Evan


  • TSL
    31 January 2012

    great article on the complexity of cancer!


  • TSL
    31 January 2012

    great article on the complexity of cancer!