Diagnostics and personalised medicine – the next 40 years

As personalised medicine moves ever closer, we look at what’s changed in diagnostics and what the next 40 years hold During the last century we have seen enormous advances in medical diagnostics across a growing number of disease areas, ranging from rapid high volume screening of blood and urine through to improved microscopy and culture techniques, and more recently to the ability to identify and screen for genetic diseases. However, advances in diagnostic technology and the rapid progress in personalised (or stratified) medicine means that the next 40 years is likely to be transformational in terms of the speed, accuracy and availability of new diagnostic technologies, with the concomitant improvement in human health.

[caption id="attachment_24816" align="alignright" width="300" caption="Table 1"][/caption]

Historically, the earliest aim for diagnostics was to establish the nature of a patient’s illness, i.e. classical diagnosis. However, not only are we able to now rapidly diagnose many diseases, but we are also able to give a much more accurate picture of that disease. Figure 1 shows an example of the progress in the area of haematology. 100 years ago it was possible to identify that a patient had a blood disease by simple microscopic examination of the morphology and distribution of cell types in the blood. As time has progressed and new discoveries made and new technologies developed, it is now possible to differentiate leukaemias and lymphomas into almost 90 specific classes. Most importantly, this ability to sub-classify the level of disease has resulted in more directed and specific means of treating these diseases and allowed pharmaceutical companies to more accurately target the individual pathologies. The outcome of this has been that five-year survival rates for patients with haematological cancers have increased markedly over the past century.

Apart from the ability to sub-classify disease, the advent of more sophisticated technologies in molecular diagnostics has also broadened the utility of diagnostic procedures to cover other aspects of medicine. Figure 2 illustrates that it is now possible to radically change the way the classical medical paradigm operates, to a point where generic treatment is increasingly being replaced by truly personalised treatment of the individual. For example, it is now possible to determine a patient’s risk of susceptibility to a particular disease e.g. the use of BRCA1/2 screening in breast cancer. Such biomarkers may also be able to give a more accurate estimate of disease prognosis and to show whether a patient is responding to therapy or whether their disease is progressing. More recently, this has taken a major step forward with the ability to specifically identify those patients who are likely to respond to a particular drug and those who will not receive any benefit.

It is now possible to radically change the way the classical medical paradigm operates, to a point where generic treatment is increasingly being replaced by truly personalised treatment of the individual

[caption id="attachment_24820" align="alignleft" width="300" caption="Figure 1"][/caption]

Tailoring a patient’s treatment or drug selection to their disease is one of the most important and fastest growing areas in medicine today. Statistics show that in many areas the rates of efficacy for classical therapeutics are extremely variable, particularly in the area of oncology where, until recently, only one in four patients were likely to derive any therapeutic benefit from the drugs that were used. Not only does this mean that there are large costs to the healthcare providers for using the often expensive drugs which may have absolutely no benefit for the patient, but also that the patient is being exposed to trial-and-error medicine and their condition may not be controlled or cured as efficiently as it might if the correct drug at the correct dosage level was used.

Apart from the lack of efficacy in some therapeutic areas, it is also clear that many patients are being treated with drugs that are not only ineffective but are also causing significant side-effects. Figure 4 illustrates the scale of the problem in terms of adverse drug-related events in both Europe and North America.  Again, this causes major issues for healthcare providers and patients alike. Patients are being prescribed drugs that are ineffective for them but, also, the adverse side-effects associated with those drugs could be causing additional problems. For the healthcare providers, the inefficient use of drugs has already been mentioned but, additionally, the side-effects result in more cases of hospitalisation, additional investigations, outpatient management and huge amounts of associated costs.

[caption id="attachment_24821" align="alignright" width="300" caption="Figure 2"][/caption]

These two areas, lack of efficacy and the associated side-effects, have been the main drivers for personalised medicine and this is one of the fastest growing areas in medicine. Classically it has been known for many years that one-size-fits-all generic medicine is an inefficient approach to patient management, but the tools have not been readily available to do anything about it. Recently this situation has changed radically, particularly in the areas of virology and oncology. In virology, many of the foundation stones of personalised medicines were laid down with the recognition that viruses can respond to drugs by mutating their replicative enzymes, so that they become resistant to the therapy. Consequently an HIV patient (for example) that is treated with a specific drug for six months may no longer be responsive to that drug, and a new drug needs to be introduced. Techniques of molecular genotyping of the virus have therefore been developed to routinely monitor patients for the appearance of these mutations, so that the drugs can be changed as soon as the mutations appear.

Classically it has been known for many years that one-size-fits-all generic medicine is an inefficient approach to patient management, but the tools have not been readily available to do anything about it

[caption id="attachment_24824" align="alignleft" width="300" caption="Figure 3"][/caption]In oncology, the classic example has been the introduction of herceptin as a therapy for breast cancer but only in patients whose tumours possess the Her-2 biomarker on their surfaces as an indication that the drug target is actually present, and that the tumour is likely to respond to the therapy. More recently, other oncology drugs have followed similar routes as a result of observations that in clinical trials only specific cohorts of patients have responded to particular drugs. For example, in colorectal cancer two new therapeutic antibodies, Panitumumab and Cetuximab (see Figure 5) have been shown to be effective only in patients whose tumours possess the wild-type sequence of the target protein (k-ras), whereas any tumours whose k-ras sequence is mutated are non-responsive. Similarly, it has been clearly demonstrated that Gefitinib, a drug for non-small-cell lung cancer (NSCLC), only works in patients whose tumours have activating mutations in the target protein, the EGFR receptor.

The consequence of these findings has been that these drugs have been licensed for prescription only to those patients whose tumours have been genotyped to analyse the nature of their targets. The new concept of companion diagnostics has arisen since these drugs require the companion assay to be available commercially at the same time as the drug is licensed. This a major step forward in personalised medicine, because it means that patients can now be treated  with a drug that will actually work for them, and if they are not suitable for that therapeutic then an alternative treatment can be used.

The availability of the tools to identify and profile these tumours is a complete shift in medical practice that has enormous implications throughout healthcare. Firstly, the pharmaceutical companies are able to more accurately design their clinical trials and can stratify cohorts within the trials so that they can identify those patients most likely to gain therapeutic benefit. By introducing such stratification into trials the companies are able to reduce the numbers of patients required, the speed of the trial (usually) and importantly the overall cost of the trial. Secondly, drug-regulating authorities such as the FDA and EMEA are now issuing guidance that in the future, where possible, new therapeutics will be required to have associated companion diagnostics biomarker data available as part of the drug submission dossier. Exceptions are only likely to be those drugs which address serious life-threatening disease and for which it is unlikely that relevant biomarkers can be easily identified. Thirdly, it means that clinical diagnostics are beginning to adopt a far more important role in medicine; in fact the availability of an appropriately developed and validated diagnostic test may be the rate-limiting step in introducing a new drug onto the market.

So, given the growing importance of biomarkers in personalised medicine, what can we expect to see change in the next 40 years?  Firstly, to return to the cancer situation, the identification of individual biomarkers such as k-ras or EGFR has been a radical breakthrough in managing colorectal cancer and NSCLC. Unfortunately, further clinical analysis still shows that some patients do not respond to these drugs as expected, even though their tumour profiles are suitable. In these patients, further investigation has now shown that other components of the target biochemical pathway (see Figure 5) may also be mutated, and if that is the case, then the patient may still not respond. What this indicates is that, for an individual tumour, future drug prescription to identify efficacy may require analysis of multiple factors simultaneously. For example, a drug targeting the EGFR pathway might require analysis of k- ras, b-raf, MEK and MAPK to ensure that all components of the target pathway are suitable for that drug. Therefore we are likely to see new diagnostic platforms emerge that are capable of determining multiple biomarkers at the same time, in contrast to the current situation where molecular analysis of multiple mutations, e.g. k-ras or EGFR, are usually monitored separately for each individual gene rather in combinations. Already new platform technologies are being developed to enable this multiplex analysis, and these platforms are being increasingly used in centralised reference laboratories to deliver this information.

As the level of genomic diagnostic complexity increases the technology required to deliver this analysis also needs to evolve and a major development which is already being seen to emerge in clinical diagnostics is the use of whole genome sequencing (Next Generation Sequencing – NGS). This is probably the most powerful emerging diagnostic tool, and with the technology improvements and cost reductions it is genuinely possible to sequence a patient’s entire genome sequence at an affordable cost. As this is the ultimate multiplex analysis, it is going to be an extremely powerful tool for profiling a patient, whether from diseased tissue or as a normal baseline. There may even be a move towards introducing NGS analysis as part of neonatal screening so that an individual patient profile can be determined, incorporated into medical records and used for pharmacogenetic analysis and clinical management throughout the patient’s life. Ultimately, this technology delivers so much information that analytical tools will need to be established to filter relevant information, so that these multiple biomarker profiles can be analysed nearer  to the patient and ultimately at the point-of-care, if not in the home. Already biosensor technologies are emerging which look capable of delivering such information particularly for protein-based biomarkers. For nucleic acid-based analysis, the technical challenges have been to produce PCR-based technologies that can deliver a result fast enough to be delivered in a doctor’s office or emergency setting. Recent developments from several companies using ultrafast heat-exchange, isothermal cycling and nanotechnologies, for example, are now capable of producing monoplex or low multiplex (<5) analyte measurements within 30 minutes using hardware no bigger than a laptop computer. It is likely that the next significant development in this area will be the extension of such technologies into genuine multiplex analysis tools for detailed patient profiling.

Over-riding all of the new technical advancements that are emerging now are the analytical tools used to analyse, interpret and report this vast amount of biomarker information for rapid clinical decision-making

Over-riding all of the new technical advancements that are emerging now are the analytical tools used to analyse, interpret and report this vast amount of biomarker information for rapid clinical decision-making. During the next 40 years, the use of sophisticated bioinformatic systems including cloud computing will evolve into a major component of clinical management as we see radical changes in personalised medicine and molecular diagnostics. Although this focus has been largely on molecular diagnostic analysis, there are also likely to be major developments in miniaturisation of technology, in vivo imaging and telemedicine which are also going to transform modern medicine and make major improvements in patient healthcare.  In the next 40 years we are likely to see changes in diagnostic technologies that will truly revolutionise personalised medicine.

By Berwyn Clarke, Chief Scientific Officer, Lab21