Doctors have traditionally approached diseases with a one-size-fits-all model. But advances in genetics and artificial intelligence are making medical treatment far more personal and effective. Matthew Partridge explains how to profit from this medical revolution.
Medical treatment has followed the same broad pattern for thousands of years. You wait until you get sick, then you go to the doctor, who tries to diagnose what’s wrong. Once they work it out, they give you what is deemed the correct treatment for the condition. But while this model can work, advances in medical science have exposed its limitations.
Not only do symptoms take a long time to emerge, but diagnoses can be inaccurate, so by the time people go to the doctor, the condition may be incurable.
Another problem is that not all drugs and treatments are as effective for some people as they are for others, even when treating the same condition. If you are one of those lucky people who respond well to the standard treatment, then all is well. But if you aren’t, then the wrong cocktail of drugs could mean the difference between a full recovery and continued illness, or even death.
A new era is dawning
But things are changing. The past few years have seen huge advances in medicine, heralding a new era in treating disease. We’re getting better at predicting diseases, especially those that run in families, which means that both doctors and patients know which symptoms to watch out for. We’re more effective at tailoring treatments to the individual, resulting in healthier patients and fewer side effects from medication, says Ketan Patel, manager of the Amity UK Fund at EdenTree Investment Management. Both sets of advances could have the knock-on effect of “helping to alleviate financial pressures on healthcare systems that are already under considerable strain”. This revolution is down to advances in three fields: genomics, epigenetics and artificial intelligence.
It’s in the genes…
Although we’ve known about DNA since the 1960s, it wasn’t until the completion of the human genome project in 2003 that scientists were able to map the entire human genome: all the hereditary information encoded in DNA. This led to “a lot of hype around the potential impact of genomics” (the combination of genetics and medicine), says Frances Salisbury, a European patent attorney and partner at intellectual property law firm Mewburn Ellis. Hopes that this new knowledge would enable us immediately to unlock cures for every disease proved unfounded, with drug companies and government pouring huge sums of money into this area, only to become frustrated by “just how complicated biology is”.
Still, thanks in part to the huge investments carried out during and after the human genome project, “we’ve come a long way since then, both in terms of the speed of getting genomic information and the detail that can be obtained”, says Salisbury.
One indication of how fast the technology is progressing comes from the cost of sequencing (mapping) a person’s genome. The original project cost $2.7bn and took 13 years to complete. By 2006, the cost of mapping a person’s individual genome had fallen to $300,000. Companies now offer people the chance to map their entire genome for $1,000 or less.
If mapping one’s personal genome is starting to move into the realm of affordability, then tests for individual genes are even cheaper. In 2013 the actress Angelina Jolie made headlines when she announced that she had decided to have pre-emptive surgery because the results of a test for a gene associated with breast cancer indicated that she was at high risk of developing the disease. With the costs of such tests now around £175, it has become routine for women with a family history of breast cancer – and men whose fathers had prostate cancer – to get tests and then decide how to minimise their chance of developing the disease.
… but will it emerge?
Despite recent advances, “much could be done to improve the accuracy and precision of predictive medicine”, says Andy Lindstrom of drug development firm Sygnature Discovery. In particular, “while we are good at giving people an estimate of their potential risk compared to the rest of the population, we are
still not able to give a definitive answer to the question of whether they will get the disease or not”. As a result, many people will end up undergoing unnecessary surgery, with all the risk that involves. Still, improvements in our knowledge in this area mean that “for many conditions we should soon be able to cut out this problem and give people a fully informed choice”.
Coming up with a better indication of if and when people will get certain conditions can be extremely useful. Nevertheless, “the ultimate goal is to come up with solutions, rather than just predictions”, says Lindstrom. Fortunately, over the past few years it has become clear that there are “big opportunities in using genetic information to deliver the right treatment to the right patient at the right time”. In particular, trials indicate that many drugs licensed for one particular use “can be extremely effective for other conditions, provided they are given to the person with the correct genetic profile”.
Applying drugs to different diseases
One notable example that emerged from a recent large-scale trial of personalised medicine involved a woman who had failed to respond to any of the recommended treatments for her cancer. When she was dispatched to a hospice with just two weeks to live, her doctors decided to make one last attempt to save her by switching to another drug, not normally used, based on her genes.
To their surprise, she recovered so much that she was able to leave the hospice and go back home to spend Christmas with her grandchildren. Overall, around 10%-40% of patients with apparently untreatable conditions could profit from switching to different drugs selected as part of a personalised approach, reckons Lindstrom.
A telling illustration of how much attention the scientific community is paying to genomics is that the British government is funding the 100,000 Genomes Project. Run by the NHS, the study’s aim is to map the genome of 100,000 people and then make comparisons with their past and present medical history. The hope is that this project, which is mirrored in several European countries, will make it easier to understand which specific genes affect which conditions, improving the accuracy of predictions and facilitating effective treatments, especially for rare diseases and cancers.
The private sector is also investing large sums, says Alex Hunter, global equities analyst at asset management group Sarasin & Partners. For example, biotech company Amgen has set up a subsidiary called deCode. It “has benefited from the examination of the interplay between 100,000 Icelandic patients’ genetic information and their medical records”.
Epigenetics: how genetic code is expressed
Genes play a major role in determining our chances of getting certain diseases and the types of drugs we best respond to. But at best they provide an incomplete picture, “the medical equivalent of a black-and-white photo, compared with a three-dimensional scan in full colour”, says Lindstrom. Scientists are increasingly realising that to get the most accurate picture we will also have to pay attention to epigenetics. If genes are the base genetic code within our cells, then epigenetics are the part of cells that determine how this code is expressed. “Just as orchestras led by different people will interpret the same piece of sheet music differently, people with the same genes may not have the same outcome if there are differences at the epigenetic level.”
Like genes, our initial epigenetic make-up is inherited from our parents. However, while genes remain relatively constant over our lifetime (which is why we can use them to trace our ancestry), our epigenetics change as we get older. They can even be altered by major events in our lives, such as disease, physical trauma, or lifestyle choices (such as diet). These changes can be passed down to subsequent generations. For example, numerous studies have shown children, and even grandchildren, of people who have undergone extreme physical hardship during wars or famine tend to be in worse health than the descendants of those who haven’t faced adversity.
The fact that epigenetics play an important role in disease and changes over time has big implications for both predictive and personalised medicine, says Tom Stubbs, chief executive of epigenetic company Chronomics. Because epigenetics “can shine a light on where our health is heading years before we even start to have any meaningful symptoms”, we can “find out about our epigenetics today and take action to avoid ill health tomorrow”. It also enables people to cut through all the contradictory health advice they are bombarded with.
The field of epigenetics “is moving faster than ever, with new technologies and breakthroughs in our understanding occurring almost daily”, says Stubbs. We will soon see “huge gains in our understanding of epigenetics and an ever-broadening set of indicators that can tackle ill health proactively”. For example, Stubbs’ company Chronomics has developed a non-invasive saliva test to help patients “understand how to avoid many of the largest risk factors associated with age-related conditions, such as heart disease and cancer”. These include factors that were previously invisible, “such as smoke exposure, metabolic status and biological [not just chronological] age”.
How artificial intelligence can help
Personalised medicine has been greatly advanced by developments in human biology. But artificial intelligence (AI) is playing an important role too, not least by sifting through the huge amount of evidence generated by genetic and epigenetic research. “AI and deep learning can identify meaningful patterns and relationships in raw data that would take humans years… to recognise,” says Marcus Vass, a digital health specialist at legal practice Osborne Clarke. It can then turn these patterns “into much earlier prediction of which patients will develop which specific disease”.
While genetic testing will contribute to the “exponential” growth in information available for diagnosis and drug development, there’s also “the digitisation of health records and the internet of medical things”, as Vass points out. Healthcare providers will also be able “to combine new patient-generated data streams such as wearables [the data generated live by gadgets such as Fitbits and Apple Watches] and new patient-reported data streams such as mood-logging with provider data such as radiology images and lab results”. Overall, the total amount of medical information is expected to reach 25,000 petabytes by next year. One petabyte is equivalent to one million gigabytes, or 1,000 desktop PC hard drives.
Already AI systems make surprisingly accurate predictions about patients. Technology company FDNA has developed an app called Face2Gene, “which uses deep learning AI software, DeepGestalt, to spot facial features often associated with rare genetic disorders”, says Vass. Its database contains information on more than 10,000 diseases. DeepGestalt claims “to be able to distinguish between pictures of patients with one syndrome [and] another random syndrome with an accuracy rate of over 90%, beating expert clinicians by a margin of 20%”. The University of Pennsylvania has been working with Intel to create a platform called Penn Signals. Their algorithms help predict and prevent sepsis and heart failure. The platform claims to be able to identify “85% of sepsis cases 30 hours before the onset of septic shock”. Traditional identification methods can only discern septic shock two hours before it occurs.
The fact that machines “are better at pattern recognition than interpolating data” means they will always be better at diagnosis and prediction than treatment, says Hunter. Still, in the longer run the hope is that computers and AI will be able to speed up the development of personalised treatments. That would be good for producers too: “it would make drug development more targeted [and] cost effective to the drugs industry”. We are, in short, on the threshold of revolutionary change for patients and drug producers alike. We look at firms and funds poised to profit from the new era of bespoke medicine below.
What to buy now
Start with Amgen (Nasdaq: AMGN), one of the largest biotechnology companies in the world. It “has been at the forefront of cutting-edge biotechnology since it was established in 1980 and has more recently been using data analysis to help develop therapies”, says Alex Hunter of Sarasin & Partners.
It is working with several local and national health authorities to identify the genes responsible for various conditions. Despite consistent double-digit growth in earnings per share (EPS), Amgen is relatively cheap for a US company, trading at only 12.6 times 2020 earnings, and with a solid yield of 3%.
Thermo Fisher Scientific (NYSE: TMO) specialises in diagnostic and analytical products for the medical, scientific and industrial sectors. It also produces a range of equipment to help those involved in genomic research and personalised medicine. It has developed a range of so-called next generation sequencing products that allow for faster DNA sequencing and more rapid detection of epigenetic changes.
While the stock isn’t cheap, trading at 22 times 2020 earnings, the valuation is more than justified by the fact that revenues have grown by 50% over the past five years, while EPS has increased by two-thirds.
An even riskier company is Oxford BioDyamics (LSE: OBD), a spin-out from Oxford University. Its EpiSwitch technology uses epigenetics to diagnose and predict the progression of a wide variety of conditions in individuals, ranging from Alzheimer’s to liver cancer. It also has potential applications in cutting the time and cost of discovering new drugs. At present Oxford BioDyamics is losing money. However, revenue is starting to take off, thanks to fruitful research projects, and it is planning to expand further in overseas markets, including Asia. With major drug companies expressing an interest in its technology, there is a large potential upside for its shares.
An investment trust worth considering is Syncona (LSE: SYNC). It specialises in life sciences companies, and many of its investments “are at the forefront of personalised medicine using gene and cell therapy to design individual cures to an array of diseases from haemophilia to lung cancer”, says Hunter.
Recently, one of the companies that it helped found, Nightstar, was bought from Biogen for $877m, producing a return of 4.5 times the initial investment. Since it was founded in October 2012 Synconca has returned 120% for investors.
One firm in Syncona’s portfolio that’s worth a look is Autolus Therapeutics (Nasdaq: AUTL). It focuses on immunotherapy, reprogramming the body’s immune system so it can identify and destroy cancer cells. It also combines its own immunotherapy treatment with other approaches, which tends to be even more effective.
Its approach is personalised in that it is not only tailored to both the specific cancer that the patient is dealing with, but also to their individual cells. While the company has no revenues at present, it expects to complete early-stage trials of four different products this year and has 76 patents. It has $188m cash on hand, which should allow it conduct research through to 2021.