A tumour is not one thing, but many. We know this because drugs that work for some patients don’t work for others and drugs that work for a while often stop working later. So cancer not only presents a lot of different targets, but targets that can evolve to escape attack.
The promise of personalised cancer medicine is to tailor treatments to the particular target each case presents. This can work better than the broad-brush approach, but only if there are reliable ways of identifying what is particular to each case and medicines adapted to that particularity. Enthusiasts for this approach claim it will transform cancer medicine to everybody’s benefit.
According to Oxford University’s Professor Tim Maughan, launching a new £5-million initiative to improve bowel cancer treatment: “We’ll identify ways to tailor treatment and ensure patients receive the drugs and other therapies that will benefit them the most, and make a significant difference to their chances of beating this common disease.”
How do we know this can work? The breast cancer drug Herceptin is one of the best examples. It improves survival in breast cancer, but only in women whose tumours overexpress a protein that encourages cell proliferation – true of about 30 per cent of patients. So Herceptin is limited to those women who test positive for the HER-receptor.
The search is now on for biomarkers in every type of cancer that can act as signposts for therapy – and the belief is that thousands exist
The HER-receptor is playing the role of a biomarker, useful to clinicians in deciding the best therapy. Some biomarkers relate to the patient’s genes, such as the breast cancer genes BRCA1 and BRCA2, which can be useful in assessing an individual’s risk of the disease. Others relate to the genes of the tumour itself.
The search is now on for biomarkers in every type of cancer that can act as signposts for therapy – and the belief is that thousands exist. In 2012 an international team profiled the genes in cell lines taken from tumours and examined any links they could find to drug sensitivity or resistance. They found hundreds.
“We studied how genetic changes in a panel of more than 600 cancer cell lines effects responses to 130 anti-cancer drugs, making it the largest study of this type to date,” said Dr Matthew Garnett of the Sanger Institute in Cambridge when the study was published in Nature. “Our key focus is to find how to use cancer therapeutics in the most effective way by correctly targeting patients who are most likely to respond to a specific therapy.”
The difficulty may be in seeing the wood for the trees. Methods for reading the entire gene sequences of tumours are now becoming affordable, but what they show is daunting. In 2012 a team from Cancer Research UK’s London laboratory sequenced the DNA from a few kidney tumours and found only a third of the mutations were shared by the whole mass of the tumour. Mutations at one end of the tumour were different from those at the other; secondary tumours that had spread elsewhere were different again. The complexity seemed insurmountable.
“I’m still quite depressed about it, if I’m honest,” team leader Charles Swanton told Nature. “And if we had higher-resolution assays, the complexity would be far worse.”
Critics such as Dr Stuart Hogarth of King’s College London say that biomarkers have been overhyped and too many are being generated with too little consistency in the way they are tested and validated.
Writing in the journal Molecular Oncology, American cancer specialists Lynn Henry and Daniel Hayes suggest he may have a point. “In spite of three decades of research and thousands of reports of circulating biomarkers, very few tumour markers have established clinical utility,” they say.
But one that has is a mutation called KRAS used in assessing whether patients with colorectal cancer will respond to treatment with Avastin.
New drugs require exhaustive and expensive trials, but many biomarkers have been generated by less exacting approaches, often using the patients who happened to present themselves. So while personalised cancer medicine is an attractive aim, rigour will be needed to ensure that its promise is not dissipated in a blizzard of biomarkers that signify little.
Cancers developing in the colon or rectum are among the most common and can often be cured by the surgical removal of the tumour, usually accompanied by chemotherapy and/or radiotherapy. When the cancer is caught early nine out of ten patients survive.
Tumours that originate in the brain are uncommon – about 12 in 100,000 people develop them every year. A brain tumour is more likely to be a secondary cancer spreading from elsewhere in the body. Outcomes vary according to the precise type of tumour.
One in nine women will develop breast cancer at some point in their lives. Regular screening can help detect the cancer early, when treatment is very effective – deaths are now at their lowest for 40 years.
Lung cancer is often detected late and hard to treat effectively. Primary lung cancers are mostly the result of smoking, but the lung is also a common site for the spread of cancers for other parts of the body. For cancers detected early, surgery can be curative.
Melanoma is the least common form of skin cancer, but by far the most serious. It can spread from the skin to other parts of the body, commonly the lungs, liver, bones, abdomen and brain. Increasing exposure to the sun is blamed. Outcomes are good when diagnosed early.
Prostate cancer is common and may develop so slowly that men die of something else before it can kill them. Outcomes are good if it is detected early, less good if it is advanced or has spread. Surgery, hormone drugs and radiotherapy are the treatments.