Scratching the surface of tomorrow's diagnostics: the Editor-in-Chief's opinion at the 15th year of Expert Review of Molecular Diagnostics

Abstract

Interview with Attila Lorincz by Claire Raison (Commissioning Editor)

To mark the beginning of the 15th year of Expert Review of Molecular Diagnostics, the journal’s Editor-in-Chief shares his expert knowledge on translational diagnostics, his opinion on recent controversies and his predictions for molecular diagnostics in 2015 and beyond.

Attila Lorincz received his doctorate from Trinity College, Dublin, Republic of Ireland, and went on to become a research fellow at the University of California, Santa Barbara, CA, USA. During Professor Lorincz’s research on human papillomavirus (HPV), he found several important and novel carcinogenic HPV types and pioneered the use of HPV DNA testing for clinical diagnostics. In 1988, Professor Lorincz’s team produced the first HPV test to be FDA-approved for patients and in 2003, for general population cervical precancer screening. Now Professor of Molecular Epidemiology at the Centre for Cancer Prevention, Queen Mary University of London, UK, he and his team are furthering translational research into DNA methylation assays for cancer risk prediction.

•	What is the most promising area of your current research?

I have started to study genome-wide methylation patterns in various cancers, including prostate, cervix and breast cancers, using deep sequencing. The goal is to better understand the way the methylome changes on the way to malignancy. This area is still very new; we are just scratching the surface of what I expect to be a lot of interesting and medically useful information.

The research is likely to have some fairly near-term applications, such as improved diagnosis and prognosis of cancers. In the longer term, it will lead to a better understanding of how current preventative medicines act in different individuals, and understanding who may benefit and who may not benefit from these medicines. I also think that it will to lead to new and improved epigenetic drugs in personalized medicine.

My main research focus is on carcinogenesis and risk for carcinogenesis, but it is important for readers to understand that epigenetics and research in DNA methylation are applicable to lots of other areas, including chronic diseases, hormonal diseases, neurological diseases, and so on.

•	What has been your most significant achievement in the past few years?

I am pleased to have developed classifiers that predict risk of cervical cancer in high-risk HPV-infected women. One of my newer classifiers is called S5 (for score 5) [in press]. It's quite promising, showing high sensitivity, with good specificity and positive predictive values for detecting precursor lesions to cervical cancer. Improved triage is becoming more important as HPV screening moves into the mainstream, there are tens of millions of women who are high-risk HPV-infected and it's a dilemma to know how best to manage them. In Europe alone, there are hundreds of thousands of HPV infected women and I think that this is an important area of diagnostics for improvement. HPV testing is a big advance and it will eventually replace cytological screening but it is a work in progress; there is not yet a complete solution. What we need is a very accurate triage classifier for women who are infected by HPV and DNA methylation is a promising test for these women.

My laboratory team and I have also developed a methylation classifier for predicting risk of death for men living with prostate cancer. This is in the early stages and a lot more work needs to be done but we are seeing something quite exciting emerging.

Particularly attractive to me as a translational diagnostician is that

for all sorts of disease. The approaches will also feature quite strongly in the coming era of personalized medicine. I have been working on DNA methylation for many years, and I have seen the results getting stronger over time.

•	What role might DNA methylation change assays play in personalized medicine?

DNA methylation and related aspects of epigenetics are likely to play an important role in personalized medicine, perhaps becoming more important than genomics per se. The reason I say this is that epigenetic mutations are far more common than genetic mutations.

In developing cancer, we see not only genetic inherited predisposition but also epigenetic inherited predisposition. As the cancer develops, we see numerous somatic mutations, which lead to the cancer becoming more aggressive. There are far more epigenetic mutations and this should help us to put together more comprehensive classifiers and more accurate predictors. Certainly, we will have to consider the DNA methylation information as part of the overall picture.

I predict that DNA methylation patterns will be used to tell us about many important predispositions. They will tell us about how our bodies change with age, the effects wrought by the environment and by our actions and our exposure to drugs and toxins. I liken DNA methylation patterns in some ways to tree rings indicating important details of summers and winters past. Similarly, DNA methylation patterns stay with people for long periods, perhaps for an entire lifetime, they are constantly changing in subtle ways. The way the patterns change may provide predictions of long-term risk of future diseases. DNA methylation information will increasingly be combined with genomic information and expression information and other lifestyle factors to provide a more complete picture of health risks and predictions going forward. We're not really properly into personalized medicine yet: we are just scratching the surface of it with a few drugs and a few diagnostic tests, but that is the future.

•	Do you think healthcare providers are going to be able to afford to provide personalized medicine?

At the moment, this is only possible in a limited way. An important issue is affordability. Manufacturers need to provide better and cheaper ways of producing tests and drugs. The regulatory agencies need to improve the approval processes to reduce the regulatory costs, which at the moment are very expensive and have been limiting progress.

As a society we have done relatively poorly in terms of efficiency in translating research discoveries to the bedside over the past few decades compared to 30 or 40 years ago. There is an increasing tendency to slowness of adoption. The realization of a future where people can mostly avoid diseases, such as cancers or Alzheimer’s, perhaps by having early tests and preventative drugs, are quite far in the future on our current track. Even with all the excitement of genome-wide sequencing and the genomics revolution, things are happening slowly.

•	What has changed in the last 30–40 years to make us currently appear to be going so slowly?

Funding for research in general but particularly translational research has not kept pace with needs. The careers of many young researchers are being neglected. It's quite difficult for young researchers to get funding and the funding levels have gone down a lot. Every year we seem to hear of less money being provided by the granting agencies and that has a chilling effect on research and career choices. It will probably drive some enthusiastic young people away and the emphasis will likely shift from what we should be doing to things that can be done more cheaply with less risk and that are overall less productive.

Certainly, I think that research leadership has been weakening in the USA and Europe and some leadership will move to other places, which is not necessarily a bad thing but it will put the historic leaders at a growing disadvantage and that needs to be taken into account.

I hope that we can get back on track soon.

•

In your review article, ‘Cancer diagnostic classifiers based on quantitative DNA methylation’ [1], you mention that a key challenge in epigenetic diagnostics will be sorting through validating the large number of potential biomarkers expected to be identified. How do you foresee this problem being overcome?

There's no easy way to do this; there is extensive work to be done, the most critical ingredient is adequate funding. The work itself is complex, but the theory underlying it is simple, so all of this can be done if there is a will to do it. What we need is to test large numbers of clinical specimens in different settings in different areas of countries across the world. Only then can we fully work out the details of the biomarker panels and understand the strengths and limitations of these biomarkers. In addition to funding, the availability of appropriate clinical specimens is also an important obstacle.

It's very difficult to get well-annotated clinical specimens that are part of archives that have been donated for these kinds of validations. The specimens that do exist are often scanty, and we need assays that use very small amounts of material. Manufacturers and the assay producers need to develop assays that use very little DNA and RNA.

It is actually easier to discover biomarkers than to validate them. Although biomarker discovery is expensive, exacting, important and exciting, it produces a vast amount of information with which we often don't know what to do. It excites researchers, but how do we bring that information to the patient’s bedside?

If we can't validate them, they don't mean very much to the patient. They have to be tested exhaustively to see which are important, how one uses the biomarkers in a quantitative way with weighting factors and which drugs they can be used with. There is no easy way to do it. There is a lot of work involved and I still don't see very much of that going on.

•	What are your thoughts on the risks and benefits of diagnosis via self-testing apps?

I think that self-testing apps are interesting and they have a role. Although potentially useful, I'm worried because anything that can be used for good can be used for bad. Self-testing apps need to be worked into routine clinical medicine. I would not like them to be used by unqualified people making inaccurate predictions and doing more harm than good. They need to be guided by professionals who understand what these things are and what they can lead to, what value they can bring and what damage they can bring. Clinicians and nurses need to get more involved and advise patients on how to keep track of their health and wellbeing every day not just when they visit the doctor’s office. In a way, the medical community has been quite slow to move on these ideas while the IT industry has been too fast to move.

We need a common road. I personally do not support commercial promotion of people sequencing their genomes and getting bits and pieces of information on unqualified markers that may say that they are at risk of a certain disease based on unvalidated alterations. It seems fine if the information is processed by a qualified genetic counselor and there are a lot of data that underpin it. However, if a given genetic change has only been shown by one or two studies across the world and merely ‘looks interesting’, it's more likely to do damage than help. It may not even be the right genetic change for that disease. I have worked in diagnostics for a long time, and there are very many false leads versus true biomarkers, this has the potential to cause a lot of negative outcomes, which will ultimately slow down progress. I've got personal experience of exactly that.

So, yes, I'm very much in favor of diagnostic apps, but this should be led by professionals and be controlled by people who understand foremost the limitations. I think that we need to focus on the bad things they can do at the same time as we focus on the good things they can do. One should always beware the vested interests, these need to be talked about right at the very beginning, at the same time as the benefits are talked about. If there is not some kind of cost–effectiveness balanced weighting, then it's premature to bring it out. People can be excitable and want to grab the next best thing, but a self-testing application isn't the same as the latest electronic gadget. What's the worst that could happen here? A person could damage their health and perhaps even die. It's that important.

•	In your opinion, how will the US Supreme Court ruling on gene patents impact molecular diagnostics?

I think that it's going to have a negative impact and I am not happy with it. It may set the diagnostics field back and one can make strong arguments that the decision was not weighed properly with respect to the chilling implications that it will have for translational medicine.

Patents played an important role in past diagnosticians’ decisions. When my company and I brought out the very first HPV test, we spent a lot of money developing the test and getting it FDA approved. Twenty-five years later that test has more than 100 competitors, the price of an HPV test has gone down dramatically, we are exactly where we need to be, long-term. HPV testing is an important addition to human healthcare. If this court decision had been there in 1980, I don't think that the HPV diagnostics story would have happened, because for about 10 years, nobody was investing in HPV diagnostics except for Digene, we put a very big effort and venture capital into the project in the belief that we would get a proper reward at the end, the HPV patents we had played an important role in the thought process.

The landscape has changed and, in my opinion, it will hurt innovative diagnostic start-ups; they will be cautious about investing money, while at this point in time, we should be seeing a lot more money coming in. Patents do have to protect the innovator, instead what has happened is that the court has devalued the research efforts and costs of finding biomarker sequences that predict disease.

•	How do you see next-generation sequencing fitting into routine diagnostics?

I think that next-generation sequencing will become more and more a routine part of the field; it's a very important technology and is going to become a central part of ‘the new diagnostics’.

Certain PCR tests are digital in the sense we look at specific nucleotides, but they're very limited because you can look at so few of them, so when we consider genome-wide and epigenome-wide changes, bioinformatics tools are central in telling us what's going on. I think it is absolutely the future. Are we there yet? No. We're quite far away. We're seeing some whole-genome sequencing for cancer patients, but the problem is that it's very expensive. The bioinformatics that are being applied are inadequate. We're mostly not seeing the big picture.

Costs have to come down 5- or 10-fold. We need additional competition, we need additional funding and we need young smart people in bioinformatics, working out the vast amount of information that is going to come. We need to be able to do whole-genome or quasi-genome-wide analysis, and we need to do it on tens of thousands of samples. At the moment, that's prohibitive and so now what you see are lots of articles that have looked at small numbers of patients, there is a big risk of what we call the overfitting problem. One finds lots of changes, and most of the ones that are found are not actually relevant to the true risks of the disease, so we're still in the needle in the haystack stage; despite the caveats, genome-wide sequencing is perhaps the most exciting thing that is going to happen over the next 10 years in diagnostics.

•	What would you like to see happen in molecular diagnostics in 2015?

This might be a pipedream, but I would like to see improved funding from governments and greater respect from lawmakers and politicians as to the importance of diagnostics. I would like to see diagnostics and drug companies put more of their profits back into research and development and less into advertising.

From a general perspective, I would like to see deep sequencing instruments and kits drop in price by a lot from where they are today, I'd also like to see comprehensive newer assays that use much less DNA/RNA so that our precious archives of donated clinical samples don't have to be exhausted.

•	What are the most pressing areas for research in diagnostics and what predictions do you have for the next 5 years in molecular diagnostics?

The most pressing areas are emerging pandemics, such as viral infections like we have recently seen with Ebola, and antimicrobial resistance, which are things that could possibly have a hugely detrimental impact on populations near term. I work in cancer and we need to keep that focus, but we are not suddenly going to see a pandemic of cancer that could kill a large additional proportion of the world’s population, wreaking havoc. In the next 5–10 years, we need to up our game on MRSA, penicillin-resistant Klebsiella and other microorganisms that are getting very common and are infecting more people.

where we have few effective drugs against infectious diseases, which are resurging in a dangerous way. We're going to contain Ebola eventually, but not before it kills tens or maybe hundreds of thousands of people, and leaves big emotional scars. What if there is a super-Ebola or another super virus that's even worse?

Similarly, antibiotics are being used irresponsibly, we are breeding superbugs and releasing them into our populations; increasingly we don't have the drugs to fight them. We are probably going to have more and perhaps bigger emergencies in the next 5 years.

Deep sequencing will become an important part of the routine medicine. It will become applied to outbreaks of viruses and bacteria to determine the epidemiological patterns of their spread.

Point-of-care devices will grow in sophistication and accuracy and these will also be important because for infectious diseases, if we can move certain point-of-care devices to the area of outbreak, we might control the infections quicker than otherwise possible.

I'm a DNA/RNA diagnostician, so my focus is really into these areas where there are major opportunities, but at the same time there are other important areas, and we need to figure out how to keep all these areas moving forward. I would not advocate reducing funding for almost any area of medicine.

for many of these areas, and as many as we can possibly afford.

Disclaimer

The opinions expressed in this interview are those of the interviewee and do not necessarily reflect the views of Expert Reviews Ltd.

Financial & competing interests disclosure

The interviewee is partially supported by the Cancer Research UK Grant C569/A10404 and has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.