Published: 16/09/2004, Volume II4, No. 5923 Page 24 25

Genetic profiling will revolutionise cancer treatment by identifying the best treatment for individual patients. That is the theory.

But how will it be put into practice? Philippa Brice reports

Technological advances and the leap in our understanding of human genetics from the human genome project are paving the way for a revolution in cancer diagnosis and management. But how far are we from realising the benefits in terms of improved patient care?

What is the technology?

DNA micro-arrays or gene chips allow simultaneous analysis of thousands of different DNA sequences. They are small solid surfaces: thousands of DNA samples are spotted on to these surfaces in a regular pattern.

Using micro-arrays it is now possible to analyse gene expression (the production of functional proteins from selected genes) in normal and cancer cells, and among different types of cancer cells.

Analysis of micro-array data can be used to study genetic changes. It can identify cancer-associated genes using patterns of gene expression that show a characteristic difference between healthy and cancerous cells.

Similarly, micro-arrays are being used to investigate different types and sub-types of tumour. Characteristic gene expression profiles (gene profiles or signatures) can be linked with different types of cancer, their behaviour and responses to treatment.

Other forms of micro-array containing tissue or protein samples are also being used to study genetics and cellular function in cancer cells.

How is it used?

The technology has already made a contribution to our understanding of the genetic changes underlying the formation and spread of cancer cells, which will ultimately contribute to the development of novel treatments. But there are more immediate applications.

Traditionally, cancers have been diagnosed based on clinical signs and microscopic examination of tumour cells. Gene profiling is now being used to identify subclasses of tumour that, although indistinguishable by normal techniques, are genetically distinct.

Such distinctions can be predictive of clinical outcome and may direct treatment options (see box).

The use of cancer genomics in diagnosis and prognosis has enormous potential to improve patient care and reduce expenditure. For example, breast cancer treatment typically involves surgical removal of the primary tumour; the majority of patients under 50 also receive adjuvant therapy (hormonal or drug treatment) because this increases the overall probability of survival.

However, many of these women would survive without this treatment, which is expensive and can cause adverse side effects such as secondary cancers.

Hitherto it has not been possible to identify women for whom the risks outweigh the benefits. But research suggests that by looking at the expression of a relatively small set of genes it may be possible to tell whether a particular tumour is associated with a good or poor prognosis. Patients with good prognoses are unlikely to require chemotherapy.

The prognostic value of gene profiling has been reported for various different types of cancer, including prostate, bladder and blood cancers.

For patients who require chemotherapy, evidence suggests that cancer genomics may be able to refine treatment by helping to select the most appropriate type and dose of drug for a given patient. This could improve survival rates.

DNA micro-arrays are being used to seek correlations between cancer cell profiles, drug response and clinical outcome. This 'personalised medicine' approach could reduce the adverse financial and health implications of using less effective or otherwise unsuitable medication.

What are the practical barriers?

But the potential benefits of gene profiling must be weighed against the increased costs of testing.

The high-throughput capacity of micro-arrays to analyse expression of several thousand genes simultaneously may not be required once small and reliable gene sets (probably panels of 10-100 key genes) are identified for different types of cancer.However, the lack of commercial interest in developing diagnostic and prognostic tools means this will require substantial public investment.

One of the biggest problems in confirming the clinical utility of gene profiles is that most studies using microarrays for cancer genetic profiling require validation in independent patient groups. Presently studies are difficult to compare because of differences in materials and techniques used for experiments and analysis.

There are moves to standardise collection and storage of samples and data from micro-array experiments and increase availability of data to other researchers.

The UK National Cancer Research Institute has launched a UK tumour bank - the National Cancer Tissue Resource - funded by the Department of Health, Cancer Research UK and the Medical Research Council. This will link a network of tissue banks and clinical trials with processing centres where microarray data can be generated and stored.

Once gene profiles are successfully validated for a particular cancer, large-scale, long-term clinical studies will be required to confirm the reliability of their use.

Such studies will take several years to complete.

What needs to be done?

The steps needed to translate research into clinical practice include:

funding for research and development into cancer genomics;

the setting up of an infrastructure for storage and use of tissue samples and associated clinical data, to include the establishment of the national cancer tissue resource;

modification of the human tissue bill to provide legal protection for those using tissue for testing and research;

improvements to NHS IT to cope with the projected increase in data from clinical research and practice;

establishment of a formal evaluation framework for novel diagnostic techniques, similar to that provided by the Medicines and Healthcare Products Regulatory Agency for novel therapeutics;

research into the ethical, legal, economic and social implications of this emerging field.

Routine use of genomic analysis for managing some cancers could be only a few years away, but its wider application will require substantial public investment.

The NHS needs to keep apace of technological change and anticipate developments in clinical care if it is to realise the benefits of cancer genomics.

Further information

2004 cancer genomics workshop report. www. phgu. org. uk

National Cancer Resource information. www. ntrac. org. uk

Cleator S, Ashworth A.Molecular profiling of breast cancer: clinical implications.British Journal of Cancer.2004 90, 1120-1124. www. nature. com/bjc

Dr Philippa Brice is science policy and dissemination manager at the Public Health Genetics Unit, Cambridge.

lTo contribute articles to HSJ's clinical management section, please e-mail ann. dix@emap. com

Genetic profiling of tumours

Scientists at the Netherlands Cancer Institute in Amsterdam have found a set of 70 genes that can be used to identify two sub-groups of patients with lymph-node negative breast cancer based on clinical prognosis.

Patients with poor prognosis profiles developed secondary tumours within five years, while those with good profiles were tumour-free for the same period.Researchers concluded that for the poor prognosis group adjuvant chemotherapy was advisable even where there was no lymph node involvement at diagnosis.Further studies showed that the technique was equally effective in patients with lymph-node positive breast cancer.

The value of prognostic gene profiling using micro-arrays is to be tested in a clinical trial of 5,000 lymph nodenegative breast cancer patients.All classed as high risk will receive adjuvant chemotherapy and followed for five years to determine rates of recurrence.

The first commercial test kits for breast cancer prognosis have already been launched in Europe and the US, although their clinical value has not yet been comprehensively assessed.

Key points

Routine use of genetic profiling for managing some cancers could soon be a reality.

Its wider application in cancer care will require substantial public investment in research and development, IT and setting up a service infrastructure.

The first commercial test kits for breast cancer prognosis have already been launched in Europe and the US.